Optical fiber communication system technology has been developed extensively since the first system installation in early 1980s. Most systems are able to operate at a bit rate of 10 Gb/s with 40 Gb/s systems being prototyped and deployed. Additionally, wavelength division multiplexing (WDM) and code-division multiple accessing (CDMA) systems are being used in conjunction with high bit rates, which involves using hundreds of different wavelength channels to support overall data throughput. The ever increasing need for higher bit throughput is being driven by strong demands for transmitting all forms of data and information through global networks including real-time multimedia data, such as movies and videos.
To assist with these extremely high bit rates and multiplexing, it is often necessary to use ultra-fast pulses to achieve modulation of signals with high efficiency. Ultrafast pulse shaping can assist in forming ultra-short bit data streams to be transmitted as short bursts of light, thereby allowing an increase in the data transfer rate. Pulse shaping may also play a role in ultra-fast optical switching, filtering, and amplification.
Shaping ultrafast pulses is nontrivial as it involves working at femtosecond timescales, and dealing with effects associated with the reshaping of an optical pulse as it propagates through optical media. Pulse shaping may be achieved through a spatial Fourier transformation of an incident pulse so as to disperse different frequency components in space, and filtering the chosen frequency components selectively. A recombination of all the frequencies into a collimated beam results in the desired pulse shape.
One conventional method of pulse shaping includes using a grating to spread an input optical pulse so that each different spectral component maps onto a different spatial position. Typical conventional arrangements usually only employ phase modulation, since phase encoding produces appreciable differences in the pulse shape, and phase modulating devices such as liquid crystal modulators (LCM) are readily available. A collimating lens and grating pair are set up in a 4-F configuration (F being the focal length of the collimating lens). An element is placed at the center of the 4F system that will modulate the spectrum. For example, a spatial light modulator (SLM) may be inserted at the Fourier plane to manipulate the spectral amplitude and/or phase components of a light wave.
Microelectromechanical system (MEMS) mirror membranes have also been used for phase modulation as this technology matured, enabling production of smooth and continuous phase variation. The mirror membrane is further advantageous in that the phase modulation is not polarization dependent, although the diffraction gratings of conventional arrangements typically are. Unfortunately, the mirror membrane is more difficult to control due to the actuator coupling and the extended influence regions of each actuator which also precludes the generation of abrupt phase jumps. For both continuous MEMS mirror membrane devices and pixelated devices such as LCM, accurately setting the SLM requires a feedback control mechanism. For pulse compression applications, a nonlinear effect can be used to maximize the pulse peak power as a function of SLM settings. However, for arbitrary waveform synthesis, high-resolution pulse characterization techniques such as spectral interferometry and second harmonic generation frequency-resolved optical gating (SHG-FROG) techniques have been demonstrated for feedback control. While the latter technique does not require a reference pulse, it has direction of time ambiguity, rather low sensitivity, and strong polarization dependence.
A drawback associated with this conventional arrangement is that it tends to be polarization dependent. In laboratory experiments this is not an issue, however, input polarization is usually not known in fiber optic applications and may vary with time making it difficult to control phase and produce stable, accurate pulses.
Another drawback with conventional apparatus used in ultrashort pulse applications is that they tend to be bulky, and do not lend themselves, outside of a laboratory environment, to commercial implementations.
What is needed are compact, polarization-independent apparatus which are operable to produce ultra-short optical waveforms.